Logout events

Andrew,

I think that strategy makes great sense.  (And I hoped that
a bright 
JSPWiki-wise guy like you would find it pretty
straightforward.)  I for 
one will be delighted when this feature becomes available. 
It will give 
such a superior perspective on what's going on in the whole
system.

(The thing I still don't quite get is where such a listener
class 
would/could reside.  It would obviously have to be
persistent - would it 
be the engine, or a brand new class? )

Terry

PS: I also wonder would you tie the events together (to
produce the 
stats) with the login name?  Or would you (as I commented
earlier) 
really need to somehow get the events to become aware of the
session id?


Andrew Jaquith wrote:
> Terry and all --
>
> Session expiration events are an excellent idea!
There's very little 
> we would need to do here, because we already have a
session monitor 
> thread (SessionMonitor) that actually detects the
expirations. It 
> would be a very simple thing to have that chunk of code
fire off this 
> new event type (expiration) when it detects that a
session has expired.
>
> As for session statistics, I've been tossing around a
few ideas in my 
> head for a while about this. Basically, I've been
thinking that it 
> would be great if we had a listener class that
subscribes to 
> login/logout/expiration events and page events and
records a bunch of 
> stats. Things that might be interesting for session
statistics:
> - Time of login, logout, and authentication change
(anonymous -> 
> asserted -> authenticated)
> - User name
> - Length of session
> - # pages read
> - # pages edited
>
> We'd need the new event type (for expirations), a
listener class 
> ("SessionStatsListener") and a class that
would contain stats for each 
> session ("SessionStats"). There's also the
log file, which could be 
> done through log4j but might be better executed through
standard file 
> operations.
>
> I don't have any plans to do this soon (not in the next
30 days, 
> anyhow) but it would be an extremely straightforward
thing to do.
>
> Andrew
>
>
> On Oct 27, 2006, at 12:06 PM, Terry Steichen wrote:
>
>> Sure.  Allows you to better set the timeout period
to minimize 
>> overhead.  Also, if you subtract the timeout period
from the duration 
>> of a session, you'll get a very good idea of just
how long your users 
>> are actually active.
>>
>> Erik Bunn wrote:
>>>
>>> On Oct 27, 2006, at 17:09 , Terry Steichen
wrote:
>>>
>>>> I suspect that most users don't use an
explicit logout, instead 
>>>> allowing the session to expire via
inactivity timeout.  Would it be 
>>>> useful to also generate an event (in
SessionMonitor) when a session 
>>>> has expired?
>>>
>>> You're probably right, and I wouldn't mind
knowing the difference 
>>> between session timeouts and logouts, either.
Makes a difference if 
>>> you have a thousand users idling away. ;)
>>>
>>> //ebu
>>>
>>>
>>> _______________________________________________
>>> Jspwiki-users mailing list
>>> Jspwiki-usersecyrd.com
>>> http://ecyrd.com/cgi-bin/mailman/listinfo/jspwiki-users
>>>
>> _______________________________________________
>> Jspwiki-users mailing list
>> Jspwiki-usersecyrd.com
>> http://ecyrd.com/cgi-bin/mailman/listinfo/jspwiki-users
>
> _______________________________________________
> Jspwiki-users mailing list
> Jspwiki-usersecyrd.com
> http://ecyrd.com/cgi-bin/mailman/listinfo/jspwiki-users
>
_______________________________________________
Jspwiki-users mailing list
Jspwiki-usersecyrd.com
http://ecyrd.com/cgi-bin/mailman/listinfo/jspwiki-users